Sentiment Analysis on Heat- and Health-Related Tweets During the Summer

Abstract

This study presents a comprehensive relationship between maximum temperature and heat- and health-related tweets, utilizing the autoregressive integrated moving average (ARIMA) models on a dataset of 12 million tweets collected from May 29 to August 27, 2016. This study selected the five cities of the USA with the highest number of tweets: Atlanta (Georgia), Chicago (Illinois), Houston (Texas), Los Angeles (California), and New York City (New York). The results reveal that maximum temperature increases are significantly associated with a greater number of negative heat-related tweets compared to positive ones, providing evidence of the emotional toll of extreme heat. The results also reveal variations in the relationship between maximum temperature and heat-related tweets, showing significant associations in Houston (Texas), Atlanta (Georgia), and Chicago (Illinois), but not in Los Angeles (California) or New York City (New York). This partially supports the need for city-specific extreme heat warning systems. However, no significant association was found between health-related tweets and heat exposure, suggesting that health-related tweets might be influenced by diverse factors beyond environmental conditions, such as socioeconomic status, demographic characteristics, and geographical locations. The findings underscore the potential of social media as a complementary tool for monitoring public sentiment and thermal comfort during heat waves. Integrating sentiment analysis into heat wave monitoring systems may provide valuable insights for mitigating the adverse effects of extreme heat.